182193 - crash opening bugtraq-digest mails via IMAP, no error message

Status: RESOLVED FIXED

(MailNews Core :: Networking: IMAP, defect)

Description

[sievers]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2a) Gecko/20020910
Build Identifier: http://ftp.mozilla.org/pub/mozilla/releases/mozilla1.2/

in the folder view, I select the Lotus-Notes (sub)folder with the bugtraq-digest
mails, and click on one in order to open it in the lower right portion of the
screen. Loading starts, and suddenly mozilla exits without any message.
Reproducible everytime, I'd gladly forward one of those mails. They have lots of
attachments...

Reproducible: Always

Steps to Reproduce:
1. open mailnews
2. enter folder with bugtraq-digest mails
3. highlight one of the mails to see it in the preview window

Actual Results:  
the mail starts to be loaded, mozilla exits without comment

Expected Results:  
Should have displayed the mail in the preview window.

Comment 1

[sievers]

Attachment: bugtraq digest, 3-Nov-2002

This is one of the mails that caused mozilla to exit without comment, hope it
is still reproducible after extracting it from IMAP via save as... Otherwise,
I'll forward one of the mails.

Comment 2

[sievers]

Some further testing shows that mozilla only crashes, if the mail is inside an
IMAP *Sub*folder. Inside my IMAP Server, I have one Folder "Lists", and if I
move one of the bugtraq-digest mails inside a subfolder (such as "Bugtraq"),
mozilla crashes on opening the mail.

Comment 3

[David :Bienvenu]

are you using a lotus notes imap server? Could you attach or send me a
client-side imap protocol log by following these instructions? Thx!

http://www.mozilla.org/quality/mailnews/mail-troubleshoot.html#imap

Comment 4

[sievers]

Attachment: imap protocol log

Here's a protocol log (I just cut out the hostname). And yes, it's a lotus
notes server that speak I speak imap to. Glancing at the log it looks ok...

Comment 5

[David :Bienvenu]

also, you could you try downloading a newer 1.3 trunk build and see if that
works for you - I'll try to reproduce this, but if I can't, there's a chance
that this has been fixed.

Comment 6

[David :Bienvenu]

I was not able to reproduce this problem with our imap server. It might be
specific to Lotus Notes servers, or it could be fixed. I don't have a lotus
notes server to test against; the one we have access to doesn't seem to be up
(or the account info I have is wrong)

Comment 7

[sievers]

I tried "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021127", same
problem. Please let me know if I can send you other debugging information!
I did remove the complete ~/.mozilla directory, set up the IMAP-Connection again
- same problem.

I had not found a similar bug report in the lists, so it might be specific to
our setup, or some weird coincidence... I'll keep looking.

Comment 8

[David :Bienvenu]

The imap protocol log looks fine. Is it possible for you to install Talkback so
we can see where you're crashing? Another possibility is to somehow get access
to a test account on your imap server so I can try this myself on the domino
server, but I'm starting to suspect that this isn't really an imap problem. It
might be specific to Linux, or your setup. Do you have any special options set
as far as viewing messages is concerned? i.e., view | message body as |
html,simple html,plain text? (I tried all three and didn't have a problem, but
it would be good to know. If you've never changed it, it would be html).

Comment 9

[sievers]

I did install everything from mozilla-i686-pc-linux-gnu-1.2-sea.tar.gz, but it
does not pop up the talkback window, when it crashes - weird.

I tried strace'ing it, and this happens towards the end. It looks like one of
the processes does find the time to kill some children on its own - if I read it
correctly, the kill is invoked by the process, not by the system.

[pid 28994] <... poll resumed> [{fd=8, events=POLLIN}], 1, 2000) = 0
[pid 28994] getppid()                   = 1
[pid 28994] kill(29002, SIGKILL)        = 0
[pid 28994] kill(28999, SIGKILL)        = 0
[pid 28994] kill(28998, SIGKILL)        = 0
[pid 28994] kill(28997, SIGKILL)        = 0
[pid 28994] kill(28996, SIGKILL)        = 0
[pid 28994] kill(28995, SIGKILL)        = 0
[pid 28994] _exit(0)                    = ?

I'm sorry, but I can not give direct access to my (or a test account), sorry.

I can reproduce the problems with a fresh install. As root, I removed ~/.mozilla
completely, installed mozilla and then set up the IMAP connection, touching
nothing else. Crash's still there.

BUT: I found out an oddity - I have created a new IMAP-Subfolder and copied the
mails to it, and tadaa - I can read the mails in the new folder, while it still
crashes in the old one.

This leads me to agree that it looks like a problem specific to this setup,
obviously not easily reproduced. The creation of a new folder seems to be a
workaround (although it still should not crash). Domino's IMAP (at least in
5.0.x) is known to be buggy, too...

I can live with the situation as is now, but if you can think of a way for me to
provide you with more information, please tell me so!

BTW, there is another bug in Domino's IMAP-Server (should be fixed R6, though).
Fetching the envelope [e.g. "a01 fetch 4 all"] on mails with more than one entry
in the "To" or "Cc" field can lead to wrong answers! That's a problem for pine.
mozilla seems to parse the headers itself ( FETCH [...] BODY[HEADER.FIELDS (From
To Cc [...])). As long as you do that, you're safe with Domino, but if you plan
on changing that into letting the imap server parse the envelope, there'll be
trouble with R5.

thanks so far!

Comment 10

[David :Bienvenu]

Can you try this with a recent 1.5 build and let me know? There have been a
number of fixes since 1.3...thx

Comment 11

[sievers]

I just tried 1.5a and could not reproduce the crash with the "broken" IMAP
folder I kept for this purpose. I tried an older version again, but it also did
not crash. Therefore, I cannot say for sure whether a change in mozilla or in
the IMAP folder (a notes fixup that doesn't trigger the bug anymore?) resolved
this issue. But, as I am now unable to reproduce the crash and if no one else is
"lurking" on this bug and speaks up, I'd consider it fixed. Thanks!

Comment 12

[David :Bienvenu]

great, thx, marking fixed.